2 level bi-factor random item respons... PreviousNext
Mplus Discussion > Multilevel Data/Complex Sample >
Message/Author
 Shanshan Wang posted on Friday, June 12, 2015 - 1:05 pm
Hi Mplus veterans!

I tried to fit a two-level item bifactor random item response model. The below are protocol scripts I wrote for the model. I will greatly appreciate if you can provide some feedback on my scripts.

Thank you!

TITLE:
Two-level item bifactor random response with categorical data


VARIABLE:
Names are
cluster y1 y2 y3 y4 y5 y6;
Usevariables are y1-y6;
CATEGORICAL ARE y1-y6;
CLUSTER IS cluster;
ANALYSIS:TYPE = TWOLEVEL RANDOM;
PROCESS = 2;
ESTIMATOR = bayes;
MODEL: %within%
s1gw-s6gw | gw by y1-y6*;
s1fw1-s3fw1 | fw1 by y1-y3*;
s1fw2-s3fw2 | fw2 by y4-y6*;
gw-fw2@1;
gw with fw1 fw2@0;
%between%
gb by y1-y6*;
fb1 by y1-y3*;
fb2 by y4-y6*;
gb-fb2@1;
gb with fb1 fb2@0;
 Bengt O. Muthen posted on Saturday, June 13, 2015 - 12:26 pm
This looks correct but may be hard to estimate given that you have random loadings for the specific factors, not only for the general factor. My input in the paper is simpler:

Model:
%within%
s1-s13 | g by u1-u13*;
verbal by u1
u5 u10 u13 (lamv1-lamv3);
person by u7
u11 u12 (lamp1-lamp2);
g@1; verbal-person@1;
g with verbal@0;
g with person@0;
! verbal with person@0;
%between%
g@0;
verbalb by u1
u5 u10 u13 (lamv1-lamv3);
personb by u7
u11 u12 (lamp1-lamp2);
verbalb personb;
! g with verbalb-personb@0;

[s1-s13] (p1-p13);
fb by u1-u13*1 (p1-p13);
sigma by s1-s13*1 (p1-p13);
fb sigma;
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: